NutriPersona: Conception of a computational tool for elaboration of personalized menu from a Brazilian food composition database

Abstract

Computational tools have been used in nutritional care. In Brazil, all diet evaluation and calculation tools use food composition data from other countries, which can influence the nutrient content of the calculated menus. Another limitation is that they do not automatically generate menus. This study aimed to develop a computational tool to represent the nutritionist’s expertise in elaborating personalized menus using the Nutrient Intake Assessment Database of the Brazilian Food Composition Table (TBCA NIA-DB). The steps for developing the NutriPersona tool included defining the necessary information during the decision-making process, restructuring the TBCA NIA-DB, and developing the algorithm to generate the menus automatically. The steps required to generate the personalized menus were organizing the TBCA NIA-DB food groups, using the finite state machine technique to implement computational strategies to elaborate nutritionally adequate menus, and considering the patient’s preferences. The tool validation was performed by generating 105 menus, which were evaluated by experienced nutritionists (n = 18) and showed 89.2% agreement on suitability for nutritional recommendations and proposed preferences. The NutriPersona tool was effective in simulating the nutritionist’s expertise; it will contribute to optimizing nutritional care and function as a decision support tool.